Neuroblastoma, the most common extra-cranial solid tumor of childhood, has poor long-term survival in the 45% of children with high-risk features at diagnosis, despite intensive therapy. Tumor specific therapy is needed to overcome resistance, since maximal dose intensity alone, even with myeloablative therapy, has only modestly improved survival. Objectives: 131I-metaiodobenzylguanidine (131I-MIBG) is similar in structure to norepinephrine and has been shown to concentrate in neuroblastoma and therefore will be tested for cell-specific radiation treatment of this tumor. Hypothesis: 131I-MIBG will produce responses in resistant tumors since it is specifically targeted to the neuroblastoma cells, and can therefore deliver much higher doses of radiation locally than conventional external beam radiotherapy. We further hypothesize that there is a relationship between delivered activity of 131I -MIBG and tumor response in resistant neuroblastoma. Previous studies have been limited by the hematopoietic toxicity and by the radiation safety limits. The dose can be augmented with a double successive infusion supported by hematopoietic stem cells (HSC). Specific Aims: (1) To determine the maximum tolerated red marrow radiation dose delivered and associated toxicities of escalating activity of 131I-MIBG infused in two consecutive treatments separated by 2 weeks when supported by HSC infusion 2 weeks after the second infusion; (2) to determine engraftment with this regimen; (3) to determine, within the confines of a Phase 1 study, tumor response; (4) to determine, in patients with evaluable lesions, radiation dose to the tumor. (5) To investigate possible correlation of TP53 mutations with response to 131I-MIBG. Design/Methods: The study will be a standard Phase I (3+3) dose escalation open to children with resistant neuroblastoma. Patients will receive 131I -MIBG on day 0 and 14, with HSC rescue on day 28. The first infusion will be based on body weight, while the second will be adjusted to reach the desired total red marrow radiation dose. Three dose levels will be studied, and a maximum of 18 patients accrued from the NANT Consortium institutions. All patients will be studied with whole body dosimetry, and, if evaluable, tumor dosimetry using conjugate planar imaging. Biologic correlations include PCR for minimal bone marrow disease and tumor TP53 mutations. The maximum tolerated red marrow dose will be defined and engraftment, dosimetry, and response correlation described. Significance: The higher tumor-specific radiation doses possible with this protocol may provide a treatment to improve response and survival in children with resistant neuroblastoma.